a content based movie recommender system for mobile application

Post on 01-Nov-2014

1.575 Views

Category:

Technology

3 Downloads

Preview:

Click to see full reader

DESCRIPTION

Final Year Project Phase 1 Presentation Slide

TRANSCRIPT

A Content-based Movie Recommender System for

Mobile ApplicationSupervisor: Dr. Khairil Imran GauthModerator: Dr. Yeoh Eng ThiamPresenter: Ahmad Arafat Bin Mohd Ali (1091105499)

- What is a Recommender System? -> Information Filtering System

- Act as an information filtering from a large collection items (such as movies) by recommending some movies to users.

- Develop a movie recommender system on a mobile phone

Project Overview:

People have problem in selecting alternative items (eg. movies)

Get movie recommended by friends or movie expert reviews from website / magazine- > Time consuming

Need large data set of movie collection Mobile Phones in the market has limited

processing power to handle the recommender system calculation. Solution - > Web Services

Problem Description

Develop an algorithm for recommending movies

Working with a large dataset of movies (source GroupLens)

Implement Web-Services (server-side) -Advantage: Increase performance in the mobile phone

Implement Mobile Application (client-side)

Project Objectives

Background Study & Literature Review

Features: -Auto Movie Recommendation-Watch List-Like Feature

Weakness: -Lack of Sign in Feature-Inaccurate Movie Ranking

The Internet Movie Database (IMDb)

Background Study & Literature Review

Feature:-Movie Rating Monitoring-Film Quizzes and Games-Video and Picture Sharing

Weakness:Rating System is not monitored

Flixster

Background Study & Literature Review

Feature:-Excellent Content Availability -Platform Varsity

Weakness:-Content Coverage is limited

NetFlix

Background Study & Literature Review

Feature:-Movie Can be shared via Facebook-Find Movie Listing using third party search engine

Weakness:- Smaller collection of movies- No rating and review features

MSN Movie

Finding of Questionnaire

50 Respondents -32 Males-18 Females

Majority of the respondents watch movie weekly

Finding of Questionnaire

Majority of the respondents have mobile phone with internet access

Finding of Questionnaire

Majority of the respondents wanted to have a movie recommender system on their mobile phone

Finding of Questionnaire

Some of the features wanted to be seen in the movie recommender system

Use Overlap Coefficient algorithm to make movie recommendation

Proposed SolutionAlgorithm

Proposed SolutionUse Case Diagram

Proposed SolutionEntity Relationship Diagram

System Interface 1

Search Function

System Interface 2

Movie Description

System Interface 3

Main page after user has logged in

Q & A

top related